Vibe Coding Liability: Who Pays When AI-Built Apps Fail?
Research shows vibe-coded apps leaking sensitive data at scale, and legal experts warn liability for AI-generated code remains unclear without human review.
Prompt-driven app creation, often called vibe coding, describes building software by describing what you want in plain language and letting an AI model generate the code, interface, and sometimes the deployment. Instead of writing every function by hand, a creator iterates through conversation: requesting features, reporting bugs, and accepting or rejecting what the model produces. The approach now extends beyond web and mobile apps into game development, where text prompts can produce playable prototypes.
The barrier to shipping software has dropped sharply. People with little or no programming background can launch working products, and experienced developers can prototype in hours rather than weeks. That speed comes with trade-offs. Code that the author does not fully understand can hide security weaknesses, and when many apps share the same generated patterns, a single flaw can repeat across countless projects. As AI-built tools move into larger organizations and public-facing services, questions are growing about accountability: who is responsible when generated software leaks data or breaks, whether that is the user, the platform, or the model provider.
The shift is also reshaping creative industries. Game engine makers and major tech companies are introducing prompt-based creation tools, and some professional developers have reacted with skepticism or open criticism. Their concerns center on quality, job security, and the value of craft.
This hub tracks:
Research shows vibe-coded apps leaking sensitive data at scale, and legal experts warn liability for AI-generated code remains unclear without human review.
Unity unveiled Spark, an AI prompt-driven game maker coming to Google's new Playground, and its Space Corridor demo drew widespread mockery.
Unity unveiled Unity Spark, a prompt-to-game tool that runs only inside Google's new Playground platform, still waitlisted and limited to users 18 and up.
Research shows AI-generated code fails security tests ~45% of the time, AI-linked CVEs are rising, and thousands of vibe-coded apps leak sensitive data.
Vibe coding reached the Navy, the dictionary and a $13.3B Lovable valuation in 2026, but studies keep finding serious security flaws in AI-built apps.